SpeechRecognitionModel.transcribe only transcribes first sentence of long 5 minute audio file

[detongz]

I have a 5 minute long audio recording, but when using huggingsound to parse it, only the first sentence was transcribed.

I'm using GreedyDecoder, like below:

import torch
from huggingsound import SpeechRecognitionModel

device = "cuda" if torch.cuda.is_available() else "cpu"
batch_size = 1
model = SpeechRecognitionModel("jonatasgrosman/wav2vec2-large-xlsr-53-english", device=device)
audio_paths = ["/path/to/sagan.mp3", "/path/to/asimov.wav"]

transcriptions = model.transcribe(audio_paths, batch_size=batch_size)

print(transcriptions)